High‐Frequency Ultrasound for Detecting Periodontal Inflammation: A Preclinical Diagnostic Accuracy Study

ABSTRACT

Aim

Ultrasonography (US) has shown accuracy in imaging healthy periodontium. This study aims to evaluate the feasibility and accuracy of US for estimating dimensions of inflamed periodontium induced by ligature and bacteria.

Methods

Periodontal tissues of maxillary as well as mandibular premolars and molars in six female mini pigs were treated with ligature and three strains of bacteria for 4–10 weeks. Before euthanization, the periodontium was imaged with US. After euthanization, cone‐beam computed tomography (CBCT) scans and histology were performed. Soft and hard tissue measurements by calibrated and masked examiners from US, CBCT, and histology were statistically compared.

Results

Seventy‐one histological samples with corresponding CBCT and US scans were available for analysis. Overall, there was a good to excellent agreement between histology and US (ICC: 0.77–0.96) for parameters such as Soft Tissue Thickness (STT), Gingival Recession, Crestal Bone Thickness (CBT), and the bone‐to‐cemento‐enamel junction (B‐CEJ) distance. However, discrepancies were observed for STT at 3 mm below the CEJ and Soft Tissue Height (STH) (ICC: 0.44 and 0.54, respectively). CBCT showed lower agreement with histology, particularly for thin CBT (< 1 mm), with an ICC of 0.20, compared to 0.90 for US vs. histology. CBCT failed to identify crestal bone in 14 cases when the crestal bone was thin. Notably, CBCT results differed more from histological measurements than US in assessing B‐CEJ and thin CBT.

Conclusion

US demonstrated substantial potential as a transformative tool for periodontal diagnostics, exhibiting high agreement with histology in determining critical parameters. Compared to CBCT, US offered advantages, particularly in cases with thin crestal bone.

This preclinical porcine study evaluated the accuracy of ultrasound (US) for estimating dimensions of inflamed periodontium. Seven periodontal parameters were measured and compared between US, cone‐beam computed tomography (CBCT), and histology. US has excellent/good agreement with histology in estimating bone level/thickness, soft tissue height/thickness at 1 mm, and gingival recession. Compared to CBCT, US has a higher accuracy in cases with thin crestal bone.

Summary.

• Background

  • ○
    Ultrasonography (US) is accurate in imaging healthy periodontium in preclinical as well as clinical studies. It is not known, though, if US can estimate dimensions of inflamed periodontium accurately since pathologic process changes the mechanical properties and dimensions of periodontal tissues.
• Added value of this study

  • ○
    Preliminary preclinical data demonstrated high agreement between US and histology/cone‐beam computed tomography (CBCT) in determining periodontal hard and soft tissue parameters, including soft tissue height/thickness, gingival recession, and crestal bone thickness/level, that are potentially useful for diagnosis and treatment planning of periodontal diseases and conditions. Especially for thin crestal bone (< 1 mm), US achieved higher accuracy, compared to CBCT.
• Clinical implications

  • ○
    Integrating ultrasonography into routine periodontal examinations could enhance diagnostics, treatment planning, and patient outcomes, reducing radiation exposure and improving personalized care.

1. Introduction

According to the 2017 World Workshop on the Classification of Periodontal and Peri‐Implant Diseases and Conditions, the periodontal phenotype includes the gingival phenotype, bone morphotype, and tooth dimensions in the context of mucogingival conditions and deformities [1]. The gingival phenotype considers elements such as the soft tissue thickness (STT) and the gingival width, while the bone morphotype pertains to the crestal bone thickness (CBT). Soft tissue height (STH) is another periodontal parameter that bears important clinical implications and is the measure of the apico‐coronal dentogingival complex including the supracrestal tissue attachment (STA) and the sulcus/probing depth, determined clinically by measuring the distance from the free gingival margin to the crestal bone [2].

Clinically, gingival phenotype categorization relies on subjective judgment [3, 4, 5, 6, 7, 8]. A common convenient practice is to place a periodontal probe in the sulcus. Depending on the visibility of the probe, the phenotype is assigned as “thin” or “thick”, with a limitation in assessing intermediate tissue thickness [9]. Other methods directly measure STT include the use of calipers, endodontic instruments, and needles [10, 11, 12]. However, these invasive methods commonly entail challenges such as improper instrument angulation, rubber stop loosening, and the need of local anesthesia [13, 14]. Furthermore, the absence of a non‐invasive technique for determining STH is emphasized in the literature, with bone sounding and direct measurement after flap reflection cited as the currently available invasive methods [15, 16, 17, 18]. As for CBT assessment, the limitations of conventional two‐dimensional dental radiographs in accurately depicting facial bone, particularly its inability to measure CBT, have been acknowledged [19, 20, 21, 22, 23, 24, 25, 26]. Cone‐beam computed tomography (CBCT) is recognized as a clinically acceptable method for CBT assessment, yet its judicious use is advised due to radiation dosage and cost considerations, especially for repeated use on the same patients [26].

The advancement of medical technology, including ultrasonography, has gained some momentum in contributing to the improvement in diagnostic sciences in dentistry [27, 28, 29, 30]. Ultrasonography utilizes sound waves to visualize internal structures non‐invasively and radiation‐free [31]. Ultrasound has shown accuracy in evaluating periodontal dimensions in ex vivo human and porcine cadaverous models [28, 32, 33, 34]. Recent studies also show clinically acceptable reliability for interpreting ultrasound images among a group of readers [33, 35]. These pieces of emerging evidence suggested ultrasound is feasible and reproducible to estimate periodontal tissue dimensions. However, a critical question remains as to in vivo ultrasound accuracy and in relation to histologic measurements, the gold standard. Moreover, the abovementioned studies only evaluate normal periodontium dimension. Periodontal tissue changes in composition, topography, and dimension due to pathologic process, such as gingivitis, periodontitis, and mucogingival defects could impact the accuracy of ultrasound. To further validate ultrasound for diagnosing periodontal diseases, it is a logic next step to measure periodontal tissue dimensions and properties that are altered due to disease. Therefore, the primary objective was to compare the three periodontal parameters, STT, STH and CBT, on ultrasound images taken on ligature/bacteria‐induced periodontium of live mini‐pigs with respective histologic measurements once they were euthanized. Additionally, CBCT were compared to the other two methods for CBT measurements.

2. Methods

This is a pre‐clinical porcine study conducted at the University of Michigan and executed in adherence to a protocol (PRO00010333) approved by the Institutional Animal Care & Use Committee. Six female mini pigs, between 6 and 18 months of age, were acquired from Sinclair Bio Resources (Auxvasse, MO, USA). This large animal model was selected for this study due to the close resemblance of their dental structure and periodontal disease manifestation to those in humans [36, 37]. The report of this manuscript adheres to The Standards for Reporting of Diagnostic Accuracy (STARD) statement.

2.1. Bacterial Preparation

Representative strains of three bacteria highly associated with periodontitis were used in this study: Porphyromonas gingivalis ATCC 33277, Treponema denticola ATCC 35405, and Fusobacterium nucleatum ATCC 25586. Porphyromonas gingivalis was grown in supplemented tryptic soy broth (TSB) or agar medium containing hemin and Vitamin K, as described previously [38]. Treponema denticola was grown in tryptone‐yeast extract‐gelatin‐volatile fatty acids‐serum (TYGVS) medium [39]. Fusobacterium nucleatum was grown in supplemented TSB broth or agar medium as described previously [40]. All strains were cultivated at 37 °C under anaerobic conditions (85% N₂, 10% H₂, 5% CO₂). Cultures were grown to mid‐logarithmic growth phase and adjusted to an optical density at 600 nm of approximately 1.0 in growth medium. Samples were prepared under anaerobic conditions, sealed, then transported to the animal facility for immediate use. These preparations were supervised by a microbiologist (CF).

2.2. Disease Induction

To initiate the study, the pigs were acclimated for one week in the animal accommodation. This was followed by the implementation of thorough supra‐ and sub‐gingival cleaning using hand instruments. In the week that followed, baseline ultrasound scans were taken for another purpose, e.g., the longitudinal blood flow and mechanical property changes unrelated to this manuscript [41, 42].

An inflammatory response was induced by subgingival placement of silk ligatures that meant to accumulate bacterial plaque in the four quadrants (maxillary/mandibular and right/left) in a staggered design. Every two weeks a quadrant was randomly chosen, in which a ligature was placed in the 3rd, 4th premolars, and 1st and 2nd molars, respectively. Ligatures were inspected and, as necessary, replaced or modified at each time point. To accelerate the pathological process, the prepared three strains of bacteria were mixed and injected in gingival tissues weekly. A volume of 5 μL was delivered at each interproximal site in the periodontal sulcus/pocket with a micropump (AL‐1000, World Precision Instruments, Sarasota, FL, USA). At Week 10, euthanasia was administered to all pigs through an intravenous injection of sodium pentobarbital (150 mg/kg) and the execution of a bilateral pneumothorax puncture. With this study design, every pig had 4‐ to 10‐week exposure of ligature plus bacterial induction to yield a variety of periodontal tissue changes.

The maxilla and mandible including the studied teeth were sectioned as a single block. These specimens were then preserved in 10% formalin for a duration of 3 days, after which they were individually sectioned in accordance with the scan locations, and subsequently placed into biopsy processing/embedding cassettes for further histological processing as detailed in succeeding sections.

2.3. Ultrasonographic Imaging Procedures

B (brightness)‐mode ultrasound images were generated employing a high‐frequency linear array probe (L30‐8, Mindray Innovation Center NA, San Jose, CA, USA) on an off‐the‐shelf scanner from the same company (ZS3). The side‐facing aperture of the array spans 16.2 mm (Figure 1). An 18 MHz center frequency was chosen, enabling second harmonic imaging through a 12 MHz transmit and 24 MHz receive (CSH 24 mode). The probe also supports virtual convex imaging, expanding the field of view. The sound intensity was governed by the scanner's auto optimization function (nominally 80%). An ultrasound standoff pad and gel (Parker Laboratories Inc., Fairfield, NJ, USA) was used for sound coupling [43].

FIGURE 1.

(a) A small form factor ultrasound transducer, coupled with a gel pad, protected by a plastic cover, was placed intraorally to obtain real‐time sagittal periodontal images. (b) The intraoral probe (Zonare L30‐8) with a side‐facing aperture measuring 16.2 mm and a central frequency of 18 MHz.

An apico‐coronally oriented notch, with approximately 3 mm long, 0.5 mm wide, and 1 mm deep, were placed on the studied tooth crown with a half‐sized diamond round bur powered by a highspeed handpiece to serve as landmarks to co‐register ultrasound images and histological slides. Sagittal views of the mid‐facial locations where the crown was marked with a notch were scanned by following the scanning protocol modified from a previous study [44]. Briefly, one of the two calibrated examiners (AS and IW) operated the ultrasound probe. Extensive training was performed by following a modified scanning protocol [45]. First, the examiner freehand swept cross the mesio‐distal width of the tooth until the notch is visible on the image. Then the probe view angle was fine adjusted by rotating the probe until the maximal notch length was seen and the soft tissue and the facial alveolar bone was approximately parallel to the probe. Finally, key anatomical landmarks were identified before saving the image, including the free gingival margin, and surfaces of the gingiva, tooth crown, root, and the alveolar bone. To prevent dimensional alterations in soft tissue due to pressure, a visible layer of ultrasound gel was maintained between the probe and the tissue surface. Ultrasound images were saved as the DICOM format and exported to the open‐source software (Horos, Horosproject.org, Nimble Co LLC, Annapolis, MD, USA) for performing the following measurements by a calibrated examiner (AS). Repeated measurements were performed with a washout period; the median relative difference is 7.3%.

Soft tissue‐related parameters (Figure 2):

• Soft tissue thickness (STT) at 1 and 3 mm below the CEJ in mm

• Supracrestal tissue height (STH) from the bone crest to the free gingival margin in mm

• Gingival recession from the free gingival margin to the CEJ in mm

FIGURE 2.

Illustrates the protocol for soft tissue measurements. A tangent was drawn to the buccal surface of gingiva from the free gingival margin. Another line forming an angle with tangent is dropped to the junction between bone crest cementum. This is soft tissue height (STH) line, and STH is measured from free gingival margin to the junction between bone crest and cementum. The angle between STH line and tangent is bisected. A rectangle is placed 1 mm from free gingival margin. One side of the rectangle is parallel to the bisector and 1 mm long. The perpendicular sides measure soft tissue thickness (STT) at 1 and 2 mm (green lines).

Hard tissue‐related parameters:

• Crestal bone thickness (CBT) at 0.5 mm below the bone crest in mm (Figure 3)

• The distance from the bone crest to the CEJ (B‐CEJ) in mm (Figure 4)

FIGURE 3.

Illustrates the protocol for measuring crestal bone thickness. Using bone/root intersection as reference a rectangle is aligned with root direction. The coronal most corner of rectangle is coronal part of bone (“intersection” of crestal bone and root). The rectangle is 0.5 mm tall (H) in apical‐coronal direction (which means 0.5 mm below the crest). Box width (W) is adjusted such that box corner diagonally from bone root intersection corner is touching outer bone surface. The apical width of box (in green) is the crestal bone thickness.

FIGURE 4.

Illustrates the how gingival recession was measured. The yellow dotted line drawn from the FGM to the CEJ was the measurement of recession. The blue dotted line measures the distance from CEJ to crestal bone.

2.4. CBCT Measurements

The acquired specimens underwent scanning using a CBCT scanner (3D Accuitomo 170, JMorita, Japan) before histologic processing. The standard scanning parameters were employed, yielding an isotropic spatial resolution of 270 μm with a field of view (FOV) of 170 mm x 120 mm. The resultant CBCT slices were three‐dimensionally reconstructed using built‐in software and saved in the DICOM file format. Another calibrated examiner (JM) conducted CBT and B‐CEJ linear measurements on the selected slice that best matched the recorded ultrasound images based on the placed notches and tooth orientation.

2.5. Histomorphometric Evaluation

Collection of histological samples was facilitated using the previously mentioned notches as points of reference after euthanization. Bucco‐lingual/palatal sagittal sections encompassing the examined teeth and surrounding periodontium were obtained, demineralized in a 10% solution of EDTA, adjusted to a pH of 7.2–7.4, until they were ready for histological processing. Subsequently, all samples were transitioned to a 70% isopropyl alcohol solution. These prepared samples were then sent to the Histology Core at the University of Michigan School of Dentistry for dehydration, embedding in paraffin blocks, and sectioning into slices of 5 μm thickness followed by standard H&E staining. Once prepared, the slides were digitized at a resolution of 1 μm per pixel (equivalent to 4× magnification) using microscope (Nikon Eclipse TiE inverted microscope, Nikon Instruments Inc., Melville, NY, USA). A single calibrated examiner (JM) conducted linear measurements of the five abovementioned parameters on the histology slides using 3D‐Slicer [46] (version 5, http://www.slicer.org/).

2.6. Statistical Methods

A software package was used to conduct statistical analyses (SPSS Statistics for Windows Version 22.0, Armonk, NY, USA). Descriptive analysis was performed to calculate the mean absolute differences between the three methods under consideration, namely, ultrasound, histology, and CBCT. Furthermore, we estimated the intraclass correlation coefficient (ICC) among the three methods using mixed linear regression models. Specifically, ICC values below 0.5 indicated poor reliability, values between 0.5 and 0.75 suggested moderate reliability, values between 0.75 and 0.9 indicated good reliability, and values > 0.90 were indicative of excellent reliability [47]. Bland–Altman plots were used to demonstrate the difference distributions between ultrasound and histology, as well as CBCT and histology when the histologic crestal bone thickness was < 1 mm compared to ≥ 1 mm. All the statistical analyses were carried out with a significance level of p < 0.05.

3. Results

3.1. Sample Overview

All sites after disease induction showed clinical inflammation, characterized by swelling, erythema, and bleeding. Our recent study indicated increased blood flow in this cohort shown on ultrasound color/power flow images [48]. A total of 79 histological samples were obtained from the six pigs. Eight specimens were excluded due to complications during the histological preparation, resulting in a study sample size of 71. A representative histology slide was included in Figure S1. Histologically, accumulation of inflammatory cells (purple stains) in clusters infiltrated gingival area where the ligature was placed. Proliferations of vascular endothelial cells and accumulation of red blood cells in the lumens were evident. Proliferated and ulcerated junctional epithelium is another typical finding.

Measurements were successfully executed on all 71 samples, along with their corresponding CBCT and US scans. The analyzed samples comprised a selection of teeth: 10 from the mandibular first molar, 3 from the mandibular second molar, 12 each from the mandibular third and fourth premolars, 12 from the maxillary first molar, 12 from the maxillary third premolar, and 10 from the maxillary fourth premolar.

3.2. Parameters Related to Soft Tissue

When measured 1 mm below the CEJ, a histological examination revealed an average STT of 1.54 ± 0.48 mm, compared to the 1.66 ± 0.47 mm reading from the US examination. The absolute mean discrepancy between these two measurements was 0.29 ± 0.3 mm with ICC of 0.79 (Table 1). Three mm below the CEJ, the STT was 2.15 ± 0.54 mm histologically, in contrast to the US recorded 2.08 ± 0.86 mm. The absolute mean variance between the two values was 0.53 ± 0.68 mm with an ICC of 0.44 (Table 1).

TABLE 1.

Comparison of soft tissue related measurements using ultrasonography and histology.

Parameters	Measurements	US	Histology
STT1	Mean ± SD (mm)	1.66 ± 0.47	1.54 ± 0.48
	Mean absolute difference ± SD (mm)	0.29 ± 0.3
	ICC	0.79
STT3	Mean ± SD (mm)	2.08 ± 0.96	2.15 ± 0.54
	Mean absolute difference ± SD (mm)	0.53 ± 0.68
	ICC	0.44
STH	Mean ± SD (mm)	3.32 ± 0.65	3.03 ± 0.74
	Mean absolute difference ± SD (mm)	0.59 ± 0.59
	ICC	0.54
REC	Mean ± SD (mm)	1.78 ± 0.62	1.66 ± 0.52
	Mean absolute difference ± SD (mm)	0.35 ± 0.37
	ICC	0.77

Abbreviations: REC, recession; STT1 and 3, soft tissue thickness at 1 and 3 mm; STH, soft tissue height.

The STH, as measured histologically, was 3.03 ± 0.74 mm, as opposed to 3.32 ± 0.65 mm obtained via US. The absolute mean divergence between these two measurements was 0.59 ± 0.59 mm and the ICC was 0.54 (Table 1). Taking into consideration the average 18% tissue contraction caused by formalin fixation [49], there was an average absolute disparity of 0.33 ± 0.27 mm observed between the two measures with ICC of 0.8.

The gingival recession measured 1.66 ± 0.52 mm histologically, compared to the 1.78 ± 0.62 mm measured using US. The absolute mean difference between the two measurements was 0.35 ± 0.37 mm with an ICC of 0.77.

3.3. Parameters Pertaining to Hard Tissue

CBCT imaging could not visualize the crestal bone in 14 cases, resulting in only 57 CBT measurements. However, both histological evaluation and US assessed the crestal bone in all 71 cases. Measurements showed a CBT of 1.22 ± 0.89 mm histologically, 1.01 ± 0.61 mm via US, and 1.36 ± 0.57 mm using CBCT. The absolute mean differences were as follows: histology vs. US: 0.32 ± 0.49 mm, histology vs. CBCT: 0.3 ± 0.4 mm, and US vs. CBCT: 0.44 ± 0.35 mm. The respective ICC values were 0.83, 0.77, and 0.84 (Table 2).

TABLE 2.

Comparison of hard tissue related measurements using ultrasound (US), CBCT and histology.

CBT (excluding CBCT non‐visualized cases)	US	Histology	CBCT
Mean ± SD (mm)	1.01 ± 0.61	1.22 ± 0.89	1.36 ± 0.57
Mean absolute difference ± SD (mm) between
US and Histology	0.32 ± 0.49
US and CBCT	0.44 ± 0.35
Histology and CBCT	0.30 ± 0.40
ICC
US and Histology	0.83
US and CBCT	0.77
Histology and CBCT	0.84
CBT (Including CBCT Non‐Visualized Cases)	US	HIS	CBCT
Mean ± SD (mm)	1.01 ± 0.61	1.22 ± 0.89	1.09 ± 0.75
Mean absolute difference ± SD (mm) between
US and Histology (mm)	0.30 ± 0.49
US and CBCT (mm)	0.55 ± 0.42
Histology and CBCT (mm)	0.54 ± 0.49
ICC
US and Histology	0.83
US and CBCT	0.65
Histology and CBCT	0.75
B‐CEJ	US	Histology	CBCT
Mean ± SD (mm)	1.66 ± 0.58	1.66 ± 0.56	1.98 ± 0.73
Mean absolute difference ± SD (mm) between
US and Histology	0.18 ± 0.15
US and CBCT	0.52 ± 0.41
Histology and CBCT	0.46 ± 0.42
ICC
US and Histology	0.96
US and CBCT	0.72
Histology and CBCT	0.74

In a separate analysis, cases where CBCT scans could not detect the CBT, subsequently marked as 0 mm. The CBCT measured CBT was 1.09 ± 0.5 mm. The absolute mean differences were: histology vs. CBCT: 0.54 ± 0.49 mm and US vs. CBCT: 0.55 ± 0.42 mm. The respective ICC values were 0.75 and 0.65 (Table 2).

3.4. Accuracy of CBT Measurements Based on Thickness

Upon categorization of the cases according to histological CBT, it was observed that for CBT < 1 mm, the mean absolute discrepancy between US and histology was 0.09 ± 0.12 mm and between CBCT and histology was 0.46 ± 0.29 mm. The ICC values were 0.89 for US and histology and 0.2 for CBCT and histology (Table 3, Figure 5).

TABLE 3.

Comparative accuracy of CBCT and ultrasonography for measuring thin (< 1 mm) and thick (≥ 1 mm) bone.

Thin bone (< 1 mm)		p
Mean absolute difference ± SD (mm)
US and Histology (n = 37)	0.09 ± 0.12	< 0.001*
Histology and CBCT (n = 29)	0.41 ± 0.28
ICC
US and Histology	0.89	< 0.001
Histology and CBCT	0.20	0.067

Thick bone (≥ 1 mm)		p
Mean absolute difference ± SD (mm)
US and Histology (n = 34)	0.53 ± 0.64	0.488*
Histology and CBCT (n = 28)	0.43 ± 0.50
ICC
US and Histology	0.66	< 0.001
Histology and CBCT	0.79	< 0.001

^* p‐value of the absolute difference between US/Histology and Histology/CBCT.

FIGURE 5.

Bland–Altman Plot: assessing agreement between: (A) Histology and ultrasonography for measuring thin crestal bone (< 1 mm); (B) Histology and CBCT for measuring thin crestal bone (< 1 mm); (C) histology and ultrasonography for measuring thick crestal bone (≥ 1 mm); and (D) Histology and CBCT for measuring thick crestal bone (≥ 1 mm). Values in panels B and D, < 0.5 mm, are histology only. The corresponding CBCT values are zero.

For CBT ≥ 1 mm, the mean absolute discrepancy was 0.53 ± 0.63 mm for US and histology and 0.43 ± 0.5 mm for CBCT and histology. The ICC values were 0.66 for US and histology and 0.79 for CBCT and histology (Table 3, Figure 5).

3.5. Assessment of the Distance Between the Bone and Cemento‐Enamel Junction (B‐CEJ)

Given that the buccal bone was assessed in only 57 CBCT cases, B‐CEJ could only be evaluated in the same 57 cases. In these, B‐CEJ was found to be 1.66 ± 0.56 mm histologically, 1.66 ± 0.58 mm via US, and 1.98 ± 0.73 mm using CBCT. The mean absolute differences were: histology vs. US: 0.18 ± 0.15 mm, histology vs. CBCT: 0.46 ± 0.42 mm, and US vs. CBCT: 0.52 ± 0.41 mm. The ICC values were 0.96, 0.72, and 0.74 respectively (Table 2).

4. Discussion

In this study, the accuracy of soft‐tissue‐related parameters using US varied based on the specific parameter and location. ICCs revealed agreement of 0.79 at 1 mm below the CEJ and 0.44 at 3 mm below. The differences may arise from the mucosa's presence, with its thicker loose connective tissue that was compressible when the ultrasound probe pressured on the tissue during imaging, unlike the less elastic keratinized gingiva [50]. Dehydration from histological processing might have also altered the tissue dimension more pronouncedly in the mucosa [49]. Similar results occurred for STH comparisons between US and histology, with a 0.59 ± 0.59 mm mean difference and moderate correlation. Tissue shrinkage post formalin fixation was calculated to 16%–18% [49]. Notably, the mean difference aligns with this anticipated shrinkage, representing approximately 18% of the mean STH measured via US. This potential discrepancy should be considered when interpreting results. Moreover, the tissue undergoes some amount of distortion as soon as it is sectioned and the muscle attachment is released, as seen in the Figure S1.

One of the objectives of the current investigation was to draw a comparison between the effectiveness of CBCT and US in assessing bone levels. This was carried out by determining the distance from the alveolar bone to the CEJ. A good agreement (ICC: 0.74) between CBCT and histology, and an excellent agreement (ICC: 0.96) between US and histology suggest both CBCT and US can non‐invasively evaluate bone the level, with implications for esthetic crown lengthening and supracrestal tissue attachment evaluation.

Interestingly, in 14 instances (19.7% of total), CBCT could not visualize crestal bone, primarily due to thin bone, yet histology and US could detect it. When CBT was less than 1 mm, there was good agreement (ICC: 0.90) between US and histology, while CBCT‐histology agreement was poor (ICC: 0.2). In contrast, when CBT exceeded 1 mm, a moderate agreement existed between US and histology (ICC: 0.66), while histology and CBCT showed good agreement (ICC: 0.79). This implication is more significant when imaging peri‐implant tissues. A study by González‐Martín and colleagues revealed its inability to accurately quantify CBT around implants when it's < 1 mm thick [51]. The likelihood of buccal bone being radiographically visible exceeds 50% only when CBT surpasses 1 mm, consistent with another study indicating CBCT's inability to detect CBT when cortical bone thickness is below 0.8 mm [52].

This raises the intriguing possibility that when CBT is thin, US might offer a more reliable measurement compared to CBCT, while when CBT is thicker, the superior ability of CBCT to penetrate hard tissues may provide more accurate visualization. It is crucial to consider that a facial wall thickness of ≤ 1 mm is expected 69% of the time in the esthetic zone [53]. Consequently, the potential benefits that US might bring to this field could be substantial, especially considering these prevalent instances of thinner facial bone thicknesses in esthetically significant areas.

In the realm of therapy in periodontics and implantology, there is a growing emphasis on minimal invasiveness and preserving periodontal/peri‐implant tissue [54]. Knowing the anatomy and tissue dimensions preoperatively could potentially lead to less invasive flap designs and tissue management, mitigating potential tissue loss and esthetic complications. The existing clinical tools could be complemented by advanced imaging methods to enhance patient care [55]. Therefore, we studied US by comparing to histologic measurements, the gold standard in periodontal diagnostics. Additionally, we assessed US against CBCT for evaluating hard tissue structures, as CBCT is the primary three‐dimentional imaging method for clinical bone dimension measurements [56]. Other studies have identified additional applications and diagnostic properties of US, such as highly precise and reproducible imagistic property of US, with which bone resorption, gingival inflammation, as well as the presence or absence of subgingival calculus can be assessed ([57]). In various studies, US has also shown promising results around implants, in comparison with CBCT, μCT, and direct clinical measurements, in detecting peri‐implant bone defects [58]. US has been able to differentiate between peri‐implant health and diseased state by assessment of tissue perfusion [59].

The CEJ is a crucial landmark for identifying and managing gingival recession cases. Precise measurement of recession depth and width can be challenging when the CEJ is not discernible due to dental abrasion or cervical caries [60]. Emerging technological aids like US could prove instrumental in these challenging cases. Its potential to evaluate the surface of hard structures presented on images could bring new opportunities for precision and success in managing gingival recession where the CEJ is difficult to discern [44, 61, 62]. In the findings of the current study, there was good agreement (ICC: 0.77, mean absolute difference of 0.35 mm) observed between US and histology when evaluating gingival recession only on the facial surfaces. There was a limitation of accessibility of lingual surfaces as the pigs were intubated and had large retractors to help keep their mouths open while they were sedated. In the maxillary palate, the gingiva was very think (5–10 mm) and with thick epithelium, which affects the penetration of ultrasound wave.

Limitations of this study include the potential mucosal compression and displacement during US scan, alignment of the three imaging modalities, tissue dimensional changes during histological processing, relatively low resolution of CBCT (270 μm), and a lack of clinical standard periodontal data, such as probing depths, attachment loss, and bleeding on probing that could be used to stratify the disease severity. These limitations should be considered when interpreting the data.

5. Conclusion

In this in vivo large animal model, ultrasound‐derived periodontal soft as well as hard tissue dimensional measurements, e.g., soft tissue thickness, soft tissue height, and crestal bone thickness, aligned with histology and CBCT. Specifically, ultrasound was more accurate in delineating thin facial bone, i.e., < 1 mm than CBCT.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1. Histologically, accumulation of inflammatory cells (purple stains) in clusters infiltrated gingival area where the ligature was placed. Proliferations of vascular endothelial cells and accumulation of red blood cells in the lumens were evident. Proliferated and ulcerated junctional epithelium is another typical finding.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.